BuddhistRoad

Dynamics in Buddhist Networks in Eastern Central Asia 6th–14th Centuries

BuddhistRoad Paper 6.1 Special Issue

ANCIENT CENTRAL ASIAN NETWORKS.

RETHINKING THE INTERPLAY OF
RELIGIONS, ART AND POLITICS ACROSS
THE TARIM BASIN (5TH–10TH C.)

Edited by
ERIKA FORTE
BUDDHISTROAD PAPER
Peer reviewed
ISSN: xxx
DOI: 10.13154/rub.br.116.101
BuddhistRoad Papers are licensed under the Creative Commons—Attribution Non-
Commercial—NoDerivates 4.0 International (CC BY-NC-ND 4.0).
You can find this publication also on the BuddhistRoad project homepage:
https://buddhistroad.ceres.rub.de/en/publications/
Please quote this paper as follows:
Ciro Lo Muzio, “Brahmanical Deities in Foreign Lands: The Fate of Skanda in
Buddhist Central Asia,” BuddhistRoad Paper 6.1 Special Issue: Central Asian
Networks. Rethinking the Interplay of Religions, Art and Politics across the Tarim
Basin (5th–10th c.), ed. Erika Forte (2019): 8–43.

CONTACT:
Principal Investigator: Prof. Dr. Carmen Meinert
BuddhistRoad | Ruhr-Universität Bochum | Center for Religious Studies (CERES)
Universitätsstr. 90a | 44789 Bochum | Germany
Phone: +49 (0)234 32-21683 | Fax: +49 (0) 234/32- 14 909
Email: BuddhistRoad@rub.de | Email: carmen.meinert@rub.de
Website: https://buddhistroad.ceres.rub.de/

BuddhistRoad is a project of

SPONSORS:

This project has received funding from the European Research

Council (ERC) under the European Union’s Horizon 2020 research
and innovation programme (grant agreement No 725519).
 CONTENTS

INTRODUCTION: ANCIENT CENTRAL ASIAN NETWORKS

Erika Forte ............................................................................................ 4–7

BRAHMANICAL DEITIES IN FOREIGN LANDS:

THE FATE OF SKANDA IN BUDDHIST CENTRAL ASIA
Ciro Lo Muzio ..................................................................................... 8–43

THE EIGHT PROTECTORS OF KHOTAN RECONSIDERED:

FROM KHOTAN TO DUNHUANG
Xinjiang Rong and Lishuang Zhu ...................................................... 44–84

TIBETAN DOMINION OVER DUNHUANG

AND THE FORMATION OF A TIBETO–CHINESE COMMUNITY
Takata Tokio .................................................................................... 85–106

CONVEYING INDIA TO THE PAMIR AND FURTHER AWAY:

ON DIVINE HIERARCHY AND POLITICAL PARADIGMS
IN BUDDHIST TEXTS
Cristina Scherrer-Schaub ............................................................... 107–148
 BRAHMANICAL DEITIES IN FOREIGN LANDS: THE FATE OF
SKANDA IN BUDDHIST CENTRAL ASIA

CIRO LO MUZIO

Abstract

The relationship between Brahmanism and Buddhism is a very inspiring field of

inquiry, with which scholars have so far dealt applying different, often antithetic,
methodological approaches and attitudes of mind. The question of the ‗borrowing‘ of
Brahmanical deities by the Buddhist theistic system is one of the major issues of the
debate. On this occasion, it will be addressed from an iconographic viewpoint, with a
focus on the Brahmanical god Skanda/Kārttikeya, a multifaceted and metamorphic
deity in its own original milieu. Formerly a demon, later promoted to divine rank,
Skanda plays a role in Buddhism as well. An analysis of the diverse iconographic
contexts in which his depictions occur, from Gandhāra to Central Asia, offers inter-
esting clues to a possible explanation of his presence in the local Buddhist reper-
toires.

1. Introduction

This paper aims at presenting an overview of the representations of the

Brahmanical god Skanda in Central Asian Buddhist art, with an analysis
of his iconographic features and of the contexts in which his image oc-
curs. Following the tracks of any Brahmanical deity in Central Asian
Buddhism also means facing the question of the relationship between
Buddhism and Brahmanism, a question which cannot be reduced to an
art historical issue, as it concerns religious, philosophical, as well as
political thought.1 The topic has been matter of a lively debate, engaging
authoritative scholars and showing a tendency to polarize on conflicting
or even antithetical approaches.

____________
1
Cf. Johannes Bronkhorst‘s reassessment of the relationship between early Buddhism
and Brahmanism: Johannes Bronkhorst, Buddhism in the Shadow of Brahmanism (Leiden:
Brill, 2011).

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

11
 A major role in starting and constantly stimulating the discussion has
been played by David Seyfort Ruegg, who in 1964 formulated his own
approach to the matter and reasserted his views in the following years,
defending them from criticism in several publications covering a long-
time span2. Seyfort Ruegg interprets the presence of Brahmanical gods or
notions in Buddhism within a hierarchical frame based on the distinction
between the supramundane (Skt. lokottara) and mundane (Skt. laukika)
plans. He argues that in the laukika realm, no clear-cut separation can be
drawn between Buddhism and Brahmanism without producing artificial
and misleading results, as both systems share a common background or
‗religious substratum.‘ This is the argument Seyfort Ruegg3 opposes, for
instance, to Marie-Thérèse de Mallmann‘s view of the part Brahmanical
gods play in Tantric Buddhism, in which—as she claimed—they appear
as subjugated deities.4
In turn, the notion of a ‗religious substratum‘ is refuted by Alexis
Sanderson. Based on his analysis of the tradition of Yoginītantras, Sand-
erson argues that Buddhism and Brahmanism remain two distinct entities,
and that any similarity between them can only be explained with borrow-
ing:
The problem with this concept of a ‗religious substratum‘ or ‗common cul-
tic stock‘ is that they are by their very nature entities inferred but never
perceived. Whatever we perceive is always Śaiva or Buddhist or Vaiṣṇava,
or something else specific. Derivation from this hidden source cannot

____________
2
David Seyfort Ruegg, ―Sur les rapports entre le Bouddhisme et le ‗substrat religieux‘
indien et tibétain,‖ Journal Asiatique 252 (1964): 77–95; David Seyfort Ruegg, ―Review:
Introduction à l’iconographie du Tântrisme bouddhique by Marie-Thérèse de Mallmann,‖
Journal of the American Oriental Society 98.4 (1978): 543–545; David Seyfort Ruegg, ―A
Note on the Relationship between Buddhist and ‗Hindu‘ Divinities in Buddhist Literature
and Iconology: the Laukika/Lokottara Contrast and the Notion of an Indian ‗Religious
Substratum‘,‖ in Le parole e i marmi. Studi in onore di Raniero Gnoli nel 70° complean-
no, ed. Raffaele Torella (Rome: Istituto Italiano per l‘Africa e l‘Oriente, 2001), 735–742;
David Seyfort Ruegg, The Symbiosis of Buddhism with Brahmanism/Hinduism in South
Asia and of Buddhism with ‘Local Cults’ in Tibet and the Himalayan Region (Vienna:
Österreichische Akademie der Wissenschaften, 2008).
3
Seyfort Ruegg, ―Review,‖ 544–545.
4
Marie-Thérèse de Mallmann, Introduction à l’iconographie du Tântrisme boud-
dhique (Paris: Librairie d‘Amérique et d‘Orient Adrien Maisonneuve, 1975), 30ff.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

12
 therefore be the preferred explanation for similarities between these specif-
ic traditions unless those similarities cannot be explained in any other way. 5
The idea of a ‗religious substratum‘ or of a common theological back-
ground on the laukika level has recently received even sharper criticism
by Ronald Davidson, who labels Seyfort Ruegg‘s theorem as a meta-
physical scheme based on unacknowledged suppositions and implied
formulae, and, above all, out of time and space:
It is inconceivable, for example, that monasteries in Nāgapaṭṭiṇam [sic] in
eleventh-century South India would have had the same relationships to
‗Hindu‘ cults that were evinced by the Buddha‘s immediate disciples in
Magadha fifteen centuries before.6
In fact, the necessity of a historical perspective in retracing the interac-
tion between Buddhism and Brahmanism was a major assumption in an
essay on vaiṣṇava and śaiva motifs in Mahāyāna sūtras published more
than forty years ago by Constantin Regamey.7 On a higher level, the two
systems are distinct, yet linked by constant contacts and mutual borrow-
ings. Buddhism mirrors the changes that occurred in Brahmanism during
the time; in a sense, it provides a parallel history of Brahmanical theistic
cults.8 On the other hand, Regamey admitted the existence of a common
background shared by both systems of beliefs, though mainly in the field

____________
5
Alexis Sanderson, ―Vajrayāna: Origin and Function,‖ in Buddhism into the Year
2000: International Conference Proceedings (Bangkok: Dhammakāya Foundation, 1994),
92–93. With ―common cultic stock‖ Sanderson quotes a phrase formulated by Stephan
Beyer (Stephan Beyer, The Cult of Tārā. Magic and Ritual in Tibet, Berkeley: University
of California Press, 1973, 42).
6
Ronald Davidson, ―Review: The Symbiosis of Buddhism with Brahmanism/Hinduism
in South Asia and of Buddhism with ‘Local Cults’ in Tibet and the Himalayan Region by
David Seyfort Ruegg,‖ Journal of the American Oriental Society 129.1 (2009): 117.
7
Constantin Regamey, ―Motifs vichnouites et śivaïtes dans le Kāraṇḍavyūha,‖ in
Études tibétaines dédiées à la mémoire de Marcelle Lalou (Paris: Librairie d‘Amérique et
d‘Orient Adrien Maisonneuve, 1971), 411–432.
8
―[O]n a reconnu depuis longtemps quels texts canoniques du bouddhisme nous conser-
vent l‘image de la période de la religion populaire brahmanique intermediaire entre le
védisme et l‘hindouisme épique et pouranique sur laquelle la littérature brahmanique (à
l‘exception des parties les plus archaïques du Mahābhārata) ne nous dit rien. Les textes
brahmaniques conserves de cette époque (les Brāhmaṇa et les Upaniṣad les plus an-
ciennes) ayant un caractère élitaire sont en principe opposes aux cultes populaires sinon
franchement athées.‖ (Regamey, ―Motifs vichnouites et śivaïtes‖, 414).

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

13
of popular beliefs and minor mythological, semi-divine, and demoniac
figures.9
It is worth highlighting, in this regard, Robert DeCaroli‘s original re-
appraisal of the worship of yakṣas, nāgas and other semi-divine beings
and of the part they played in the religious mentality and everyday life of
early Buddhist monks.10 Spirit-deities are ubiquitous in the iconographic
programs of the earliest preserved Buddhist monuments as well as in
Buddhist literature, and their presence is usually explained as a compro-
mise with the beliefs of the laymen. DeCaroli rather argues that Buddhist
monks
undertook an ongoing and intentional practice whereby various sorts of
spirit-deities were transformed into Buddhist devotees. […] Images of spir-
it-deities were consistently positioned on the periphery of early monastic
complexes to reveal the new status of these beings as supporters of Bud-
dhism […].11
In this way, Buddhist monks acted as mediators between the lay com-
munity and these beings; still
they had obligations to these spirit-deities, and many members of the mo-
nastic community believed that they risked dire consequences if these re-
sponsibilities were ever ignored.12
It is evident that the relationship between Buddhism and Brahmanism is
a vast and manifold matter and that there is no univocal strategy to face
it in all its complexity. There are several possible approaches, each of
them with its own effectiveness depending on the single case or specific
text or iconographic issue we are investigating.
The imagery of Skanda in Buddhist art, and particularly in Central
Asia, represents a very inspiring case study as it summarises the terms of
the debate I have sketchily accounted for. It clearly demonstrates how
advisable a flexible approach is in interpreting the presence of deities of
ultimate Brahmanical origin in Buddhist art and cult.

____________
9
Regamey, ―Motifs vichnouites et śivaïtes,‖ 416.
10
Robert DeCaroli, Haunting the Buddha. Indian Popular Religions and the For-
mation of Buddhism (Oxford: Oxford University Press, 2004).
11
DeCaroli, Haunting the Buddha, 173.
12
DeCaroli, Haunting the Buddha, 174.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

14
Skanda is a Brahmanical god with a long history, during which he un-
derwent substantial transformations and received many names and icon-
ographic forms. Skanda is also known as Kārttikeya (as he is usually
named in Buddhist sources), Kumāra, Mahāsena; he is the much wor-
shipped Murugan or Subrahmanya of the Tamil region, which has re-
mained a stronghold of his cult, whereas in the north he receded into the
background after the 8th century13
His first representations, as a six-headed male figure, occur on
Yaudheya coins dating from the 1st to the early 2nd century.14 In Ma-
thurā and in Gandharan art (Kushan period, 1st to 3rd centuries), the god
is cuirassed, holds a spear and—in Gandhāra only—a cockerel. 15 His
classical iconography as a youth riding a peacock becomes established in
the Gupta age (ca. 4th to 6th centuries).16
The Central Asian artistic evidence of Skanda has so far been docu-
mented only in Buddhist sites of what is Xinjiang (新疆) today. Part of

____________
13For a recent insightful and innovative reappraisal of the history of Skanda, see Rich-

ard D. Mann, The Rise of Mahāsena. The Transformation of Skanda-Kārttikeya in North

India from the Kuṣāṇa to Gupta Empires (Leiden: Brill, 2012).
14
Mann, The Rise of Mahāsena, 103–107, 113–115, figs. 3a-c.
15
On the iconography of Skanda in Gandhāra, see R.C. Agrawala, ―Gandhāra Skanda
with Flames,‖ East and West 18.1–2 (1968): 163–165; R.C. Agrawala, ―Skanda with
Cock and Peacock: A Rare Device in Gandharan Art,‖ in Eastern Approaches: Essays on
Asian Art and Archaeology, ed. T. S. Maxwell (Dehli: Oxford University Press, 1992),
130–132; R.C. Agrawala, ―Skanda-Ṣaṣṭhī in Gandharan Reliefs—A Review,‖ East and
West 45 (1995): 329–332; Wladimir Zwalf, A Catalogue of the Gandhāra Sculpture in the
British Museum, 2 vols (London: The Trustees of the British Museum, 1996), I, 121–122,
II, 65–66 (nos. 102–104); Srinivasan, Doris Meth, ―Skanda/Kārttikeya in the Early Art of
the Northwest,‖ Silk Road Art and Archaeology 5 (1997–98): 233–268; Anna Filigenzi,
―Maître du ciel, héros de la terre: la triade Bouddha—Vajrapāṇi—Skanda dans l‘art du
Gandhāra,‖ in Art et archéologie des monastères gréco-bouddhiques du Nord-Ouest de
l’Inde et de l’Asie centrale. Actes du colloque international du CRPOGA, Strasbourg, 17–
18 mars 2000, ed. Zemaryalai Tarzi and Denyse Vaillancourt (Paris: De Boccard, 2005),
93–111; Mann, The Rise of Mahāsena, 123–129; Kirsten Southworth, ―Skanda in
Gandhāra: A Hindu God in a Buddhist Environment,‖ in Changing Forms and Cultural
Identity: Religious and Secular Iconographies, 1. South Asian Archaeology and Art.
Papers from the 20th conference of the European Association for South Asian Archaeology
and Art held in Vienna from 4th to 9th of July 2010, ed. Deborah Klimburg-Salter and
Linda Lojda (Turnhout: Brepols, 2014), 71–75. On Skanda in the art of Mathurā, see
Mann, The Rise of Mahāsena, 117–121 (with previous bibliography).
16
On the iconography of Skanda in Gupta sculpture and coinage, see Mann, The Rise
of Mahāsena, 203–230.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

15
the material presented in this overview—which, for reasons that will be
clear in a short while, extends beyond the borders of Central Asia—was
already discussed by Benoytosh Mukherjee.17 I will reconsider it here,
integrating it with the new interesting evidence unearthed in the Khotan
oasis during the last two decades. In Central Asia, Skanda occurs in di-
verse iconographic contexts; in spite of the variability of certain icono-
graphic details, his depictions are moulded upon the Gupta model (fig. 1).
To my knowledge, there are no Central Asian specimens of the Kushan
Skanda standing in front view, holding a spear and a cock in his hands.

Figure 1. Skanda, in a Gupta relief from Sarnath. Varanasi, Bharat Kala Bhavan
Museum. After Mann, The Rise of Mahāsena, fig. 31, courtesy Paolo Cassi.

____________
17
Benoytosh Mukherjee, ―The Development of the Iconography of Kārttikeya
(Mahāsena) in Central Asia,‖ in Investigating Indian Art. Proceedings of a Symposium on
the Development of Early Buddhist and Hindu Iconography, held at the Museum of Indian
Art, Berlin, May 1986, ed. Marianne Yaldiz and Wibke Lobo (Berlin: Museum für Indi-
sche Kunst, 1987), 249–259.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

16
 2. Skanda in Yungang

The reason why I will start my overview with Yungang (雲崗) is that
this site provides the earliest iconographic evidence of Skanda north of
Gandhāra; that the iconographic rendering of the god is particularly gen-
erous and clear; and that it shows, as we shall see, a close relationship
with the images of Skanda from Dunhuang (敦煌) and Khotan.
Skanda appears in a relief on the eastern doorjamb of the entrance to
Yungang Cave 8,18 dating from the second half of the 5th century (fig. 2).
The god is five-headed (the heads are probably intended to be six, with
one turned back, therefore not visible) and six-armed, and holds the sun
and moon discs, a bow and, very likely, arrows in his middle right hand
no longer extant, and a cockerel in the lower left hand. His lower right is
in his lap, with a small, round attribute between the thumb and the fore-
finger. His animal vehicle (Skt. vāhana) seems to merge the body and
the tail of a peacock with the head of a bird of prey, with a pearl (or a
fruit?) in its beak. A relief on the western doorjamb shows a three-
headed and eight-armed Maheśvara sitting on the bull Nandī, holding
sun and moon, a bow and arrows, a bunch of grapes, and an unclear
attribute.19

____________
18
Seichi Mizuno and Toshio Nagahiro, Yün-kang. The Buddhist Cave-temples of the
Fifth Century A.D. in North China. Detailed Report of the Archaeological Survey Carried
Out by the Mission of the Tōhō Bunka Kenkyūsho 1938–45, 18 vols (Kyoto: Jimbunk-
agaku Kenkyūsho Kyoto University), 1951–55, V, pls. 8, 11, 16–19. The authors wrongly
identify the deity as Viṣṇu (Mizuno and Nagahiro, Yün-kang, V, 98); same interpretation
in Alexander Coburn Soper, Literary Evidence for Early Buddhist Art in China (Artibus
Asiae, Supplementum, 19) (Ascona: Artibus Asiae Publishers, 1959), 238.
19
Mizuno and Nagahiro, Yün-kang, V, pls. 10, 12–15.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

17
Figure 2. Relief on the left doorjamb of Yungang Cave 8. Photo Ben Yeuda, courtesy
Rivi Ben-Yehuda.

At Yungang, Skanda and Maheśvara also appear in a relief above the

entrance to Cave 10 (last quarter of the 5th century; fig. 3). Both deities
are shown without their respective vāhana; the image of Skanda almost
completely matches the one in Cave 8 (five heads; sun and moon; bow
and arrows; unclear attribute in the right hand, in front of the chest, but
apparently not a cock), while Maheśvara is three-headed, four-armed,
holding sun and moon and an unclear attribute in the right hand, also in
front of the chest. They are portrayed at both ends of a mountainous
landscape in which worshippers and wild animals are represented.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

18
Figure 3. Relief above the entrance to Yungang Cave 10. Courtesy Huntington Ar-
chive and Huntington Digital Collection.

The prototype of the Yungang Cave 8 Skanda is the Gupta image of the
god sitting on the peacock. The presence of the cock in his hand, howev-
er, may betray a Gandharan mediation, if we consider that this attribute
is a constant feature of the Gandharan Skanda in the Kushan epoch. We
have very few examples of Skanda-on-a-peacock holding a cock, and all
of them are supposed to be post-Kushan Gandharan works.20
Apart from the Yaudheya coinage (see above), the multiple heads of
the Yungang Skanda are also rarely seen in the South Asia depictions of
the god, especially when he rides his vāhana,21 before the medieval peri-

____________
20
A relief in the British Museum, probably of post-Kushan date (Zwalf, A Catalogue
of the Gandhāra Sculpture, cat. no. 103), and a relief in the State Museum of Lucknow
(Nilakanth Purushottam Joshi, Catalogue of the Brahmanical Sculptures in the State
Museum, Lucknow (Part I), Lucknow: The Archana Printing Press, 1972, 110).
21
See, for instance, a twelfth-century sculpture from Andhra Pradesh portraying a six-
headed Skanda on a peacock, now in the Art Institute of Chicago: Pratapaditya Pal,
―Sculptures from South India in The Art Institute of Chicago,‖ Art Institute of Chicago

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

19
od. On the contrary, the association of the six heads and the peacock
matches the description of Skanda, as a member of a group of Hindu
deities (including Maheśvara, Mahākāla, Vārāhī, and others) in later
Buddhist texts, such as the Niṣpannayogāvalī.22 As to the bow and ar-
rows, they are attributes assigned to Skanda (along with a spear) in the
Agni Purāṇa and in the Matsya Purāṇa.23

3. Skanda in Dunhuang

The next piece of evidence to be considered is offered by the paintings of

Mogao Cave (Chin. Mogao ku 莫高窟) 285 at Dunhuang, dated by in-
scriptions to 538/539 (fig. 4).24 On the western wall, to the right of the
main niche housing a Buddha seated in the European fashion, there is a
triad composed by a three-faced and six-armed Maheśvara, ithyphallic
and mounting his bull, and, below him, Skanda and Gaṇeśa, in three
quarter view towards the niche. Skanda, seated on a peacock, wears a
loincloth (Skt. dhotī) and a scarf; his hair is arranged in three tufts on top
and on the sides of his head. He has four arms; in his upper left hand he
holds a trident (Skt. triśūla), in the upper right a blue lotus; in his lower
left hand, on the lap, a roundish white attribute (a cock?), in the lower
right a bunch of grapes, towards which the peacock leans its beak. A
parallel for this detail is offered by the Gupta relief in the Bharat Kala
Bhavan mentioned above (fig. 1), in which Skanda feeds his peacock

____________
Museum Studies 22.1 (1996): 31, figs. 11, 12. I thank Chiara Policardi for drawing my
attention to this artwork.
22
Benoytosh Bhattacharya, The Indian Buddhist Iconography, Mainly Based on the
Sādhanamālā and Cognate Tāntric Texts of Rituals (Calcutta: K. L. Mukhopadhyay,
1958), 363–366. Mukherjee speculates upon a possible syncretism between Skanda and
Maðjuśrī (of which we have evidence in the Mañjuśrīmūlakalpa) in the Yungang relief
(Mukherjee, ―The Development of the Iconography of Kārttikeya‖, 253–254). On this
topic, see also Seyfort Ruegg, The Symbiosis of Buddhism with Brahmanism/Hinduism,
35–36 (―Kārttikeya-Mañjuśrī in the Mañjuśrīmūlakalpa‖).
23
Mukherjee, ―The Development of the Iconography of Kārttikeya,‖ 251.
24
Paul Pelliot, Les grottes de Touen-houang. Peintures et sculptures bouddhiques des
poques des ei des T’ang et des Song (Paris: Geuthner, 1920–1924), vol. 3, pl.
CCLXV; Dunhuang wenwu yanjiusuo 敦煌文物研究所 ed., Dunhuang Mogaoku 敦煌莫
高窟 [Mogao Caves at Dunhuang], I (Beijing: Wenwu chuban she, 1982), pls. 114, 118–
119.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

20
with fruit held in his right hand.25 To the left of the niche, there is another
group of Hindu deities headed by Viṣṇu-Nārāyaṇa.

Figure 4. The gods Maheśvara, Skanda and Gaṇeśa. Dunhuang Mogao Cave 285,
western wall. After Pelliot, Les grottes de Touen-houang, 3, pl. CCLXV.

The association of Skanda with his father Maheśvara in the same Bud-
dhist iconographic frame is not a Central Asian invention. We have
iconographic evidence from Gandhāra, which, furthermore, indicates one
of the possible ways in which the presence of Brahmanical deities makes
sense in a Buddhist artistic and ritual environment. A fragment of a
much larger relief, now in the Peshawar Museum, shows a bodhisattva in
meditation (Skt. dhyānāsana) and six figures emanating from him, all

____________
25
Mann, The Rise of Mahāsena, 203. The king Kumaragupta I feeds a peacock on
coins of his ―peacock series‖ bearing on the reverse the image of Skanda on a peacock
(Mann, The Rise of Mahāsena, 219–220).

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

21
different from one another, standing on lotus flowers (fig. 5). The group
is inspired by an iconographic formula well documented in Gandharan
miracle scenes or representations of Pure Lands, with a central meditat-
ing or preaching Buddha surrounded by smaller buddhas or bodhisattvas
radiating from him.26 In this case, however, a central bodhisattva radiates
Hindu deities, and what is particularly relevant for us is that one of the
few unambiguously identifiable gods is Skanda (the central figure on the
left), shown in his typical martial Kushan iconography, i.e. cuirassed and
holding a spear. On the opposite side, on the same level, we see Mahe-
śvara, ithyphallic and holding the triśūla. Two more examples of the
same subject, in which Skanda and Maheśvara are shown according to
the same symmetric scheme, were added by Anna Maria Quagliotti: a
fragment of relief in the Chandigarh Museum27 and a whole relief in a
private collection, showing the same detail on the upper right corner,28
thus suggesting the original location of the other two fragments in the
slabs they belonged to.

____________
26
On this subject, see now Paul Harrison and Christian Luczanits, ―New Light on (and
from) the Muhammad Nari Stele,‖ in Special International Symposium on Pure Land
Buddhism, 4th August 2011, Otani University (Kyoto: Ryukoku University Research
Center for Buddhist Cultures in Asia, 2012), 69–127.
27
Anna Maria Quagliotti, ―Il Buddha che insegna la Legge: una stele raffigurante una
‗Terra Pura‘,‖ in Studi in onore di Umberto Scerrato per il suo settantacinquesimo com-
pleanno, ed. Maria Vittoria Fontana and Bruno Genito, 2 vols (Naples: Istituto Italiano per
l‘Africa e l‘Oriente, 2003), II, 641–643, 650, pls. CV, CVIII.
28
Quagliotti, ―Il Buddha che insegna la Legge,‖ pl. CV.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

22
Figure 5. Meditating Bodhisattva emanating Brahmanical deities, fragment of
Gandharan relief. Peshawar Museum. After Taddei, ―Non-Buddhist Deities,‖ fig. 1,
courtesy Peshawar Museum.

With regard to the fragment in the Peshawar Museum, Maurizio Taddei

proposed to interpret the subject in light of the ideas expounded in sever-
al Buddhist sources, among them the Saddharmapuṇḍarikasūtra and the
Kāraṇḍavyūha, according to which a bodhisattva can take many differ-
ent forms (mainly those of popular Brahmanical deities) depending on
the spiritual level and the personal beliefs of the sentient beings to be

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

23
converted to the doctrine.29 I wonder whether this explanation may suit
the Yungang and Dunhuang representations described above. At Yun-
gang, considering the spot where their images were carved (the door-
jambs of the entrance to the cave), Skanda and Maheśvara rather appear
as divine door-guardians (Skt. dvārapāla); in the composition of the
western wall of Dunhuang Mogao Cave 285, on the other hand, Skanda
features in a large composition including a number of Indian divinities
(Maheśvara, Gaṇeśa, Viṣṇu, Śakra, Sūrya, and Candra), which recalls the
enumerations, found in Buddhist texts such as the Mahāsamaya or the
Mahāmāyūrī, of Brahmanical deities who approach the Buddha and take
refuge in him.30

4. The Dandān-öiliq Evidence

We can now consider the evidence from the Khotan oasis. Three excava-
tion campaigns carried out at Dandān-öiliq (Khotan) by two distinct
archaeological missions31 led to the discovery of remarkable mural paint-
____________
29
Maurizio Taddei, ―Non-Buddhist Deities in Gandharan Art,‖ in Investigating Indian
Art. Proceedings of a Symposium on the Development of Early Buddhist and Hindu Ico-
nography, held at the Museum of Indian Art, Berlin, May 1986, ed. Marianne Yaldiz and
Wibke Lobo (Berlin: Museum für Indische Kunst, 1987), 274–278. See also, on this topic,
Seyfort Ruegg, The Symbiosis of Buddhism with Brahmanism/Hinduism, 31–34
(―Docetism in Mahāyāna Sūtras‖).
30
Seyfort Ruegg, The Symbiosis of Buddhism with Brahmanism/Hinduism, 19–29, in
particular 20.
31
a) A Sino-Swiss team led by Christoph Baumer, which carried out a three day-
exploration at the site in 1998, including the excavation of the temple D13: Christoph
Baumer, ―Dandan Oilik revisited: new findings a century later,‖ Oriental Art 49.2 (1999):
2–14; Christoph Baumer, Southern Silk Road: In the Footsteps of Sir Aurel Stein and Sven
Hedin (Bangkok: White Orchid Books, 2000); Christoph Baumer, ―Sogdian or Indian
Iconography and Religious Influences in Dandan-Uiliq: the Murals of Buddhist Temple D
13,‖ in The Art of Central Asia and the Indian Subcontinent, ed. Anupa Pande (New
Delhi: Aryan Book International, 2009), 170–184.
b) A Sino-Japanese team, which carried out diggings in 2002 and 2004, bringing to
light the temples CD4 and CD10: Yuzhong Zhang, Tao Qu, and Guorui Liu, ―A Newly
Discovered Buddhist Temple and Wall Paintings at Dandan-Uiliq in Xinjiang,‖ Journal of
Inner Asian Art and Archaeology 3 (2008): 157–170; Zhongguo Xinjiang wenwu kaogu
yanjiusuo 中国新疆考古研究所 , and Riben Fojiao daxue Niya yizhi xueshu yanjiu jigou
日本佛教大学尼亚遗址学术研究机构 ed., Dandan wulike yizhi. Zhong-Ri gongtong
kaochayanjiu baogao 丹丹乌里克遗址. 中日共同考察研究报告. Dandan Oilik Site.
Report of the Sino-Japanese Joint Expedition (Beijing: Wenwu chubanshe, 2009).

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

24
ings, which shed fresh light on the presence of deities of ultimate Hindu
origin in Buddhist Central Asia, with Skanda first and foremost. I have
already dealt with these paintings in two articles; in the first,32 I proposed
an alternative description of the divine figures depicted in the murals of
the temple D13 published by Christoph Baumer, rectifying several mis-
understandings in the reading given by the author; in the second,33 I ex-
panded the scope of iconographic comparisons and suggested an expla-
nation of the general meaning of the subjects. I refer the reader to the
latter for a complete description of the paintings, which will be given
here in a compendious form.
In the temple named D13, murals were partially preserved on the
western and northern wall of its central shrine (figs 6 and 7); on each of
the two walls (which most probably presented an uninterrupted sequence
of divine figures), three deities and a small portion of a fourth figure are
still visible. On the western wall (fig. 6), a three-headed and four-armed
Skanda, sitting on his peacock, is the first deity from the left; he holds a
bow and three arrows in the upper right and left hands, respectively, a
cock in his lower left in front of his chest, and a bunch of grapes in his
lower right hand.

____________
32
Ciro Lo Muzio, ―Culti brahmanici a Khotan: note sulle pitture del tempio D13 a
Dandan Oiliq,‖ Rivista degli Studi Orientali (n.s.) 79.1–4 (2007): 185–201.
33
Ciro Lo Muzio, ―Skanda and the Mothers in Khotanese Buddhist Painting,‖ in Inter-
action in the Himalayas and Central Asia: Processes of Transfer, Translation and Trans-
formation in Art, Archaeology, Religion and Polity, ed. Eva Allinger, Frantz Grenet,
Christian Jahoda, et al. (Vienna: Österreichische Akademie der Wissenschaften, 2017),
71–89.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

25
Figure 6. Painting on the western wall of temple D13. Dandān-öiliq. Author‘s draw-
ing after Baumer, Southern Silk Road, fig. 70.

The image reproduces the same general iconographic type found in

Yungang Cave 8 (fig. 2) and Dunhuang Mogao Cave 285 (fig. 3), that is
Skanda shown as a youth riding a peacock. With the Yungang specimen,
he shares the bow and arrows as well as the cock; with the Dunhuang
specimen, the bunch of grapes (although in Dandān-öiliq Skanda does
not appear to be feeding his vāhana). In spite of the different number of
heads (six at Yungang, one at Dunhuang, three at Dandān-öiliq) and
some mismatch in the remaining attributes, the iconographic resem-
blance is remarkable, especially if we keep in mind the presumed chron-
ological gap among Dandān-öiliq and the other two sites (three to four
centuries, according to the chronology commonly accepted for the Dan-
dān-öiliq mural painting as a whole, i.e. 7th to early 8th century).34
Even a cursory description of the other figures preserved in the D13
murals will help understand the peculiar iconographic and ritual context
in which Skanda is embedded here. The goddess to the right of Skanda,
with a swaddled child in her lap and a naked child sitting on her left leg,
can be identified as Hārītī, the small-pox bearer and children-devourer
yakṣiṇī (a semi-divine female spirit), who, after being converted to the
____________
34
Joanna Williams, ―The Iconography of Khotanese Painting,‖ East and West 23.1–2
(1973): 109–112.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

26
Buddha‘s doctrine, became a protectress of children; she was a much
revered deity in Gandharan Buddhism and was usually accompanied by
more than one child.35 With the name Hārrva, Hārītī was worshipped in
Khotan as well; she is portrayed in a well-known fragment of mural from
Farhad Beg Yailaki (6th century) with four children clinging to her.36
The third figure is a four-armed boar-headed goddess wearing a
winged crown, large round earrings, a necklace, and bracelets, holding
the sun and moon symbols, a lotus in her lower right hand, and an undis-
cernible attribute in the lower left. This deity seems to be inspired by the
Hindu goddess Vārāhī (fig. 8), the female personification Viṣṇu‘s boar
incarnation, Varāha, often appearing among the Saptamātṛkās (or
Aṣṭamātṛkās), the Seven (or Eight) Mothers.37
On the northern wall (fig. 7), the first figure on the left is Maheśvara,
three-headed, ithyphallic, holding the sun and moon discs in his upper
hands, a fruit in his lower left, his lower right resting on the knee; below,
we see his mount, the bull Nandī. Images of Maheśvara were already
recorded in Khotan, especially at Dandān-öiliq; the specimen which
appears closer to our image is the one depicted on a wooden panel from
temple DVII.6, now in the British Museum.38

____________
35
On her legend and iconographic evidence in Gandhāra, see Alfred Foucher, L’art
gréco-bouddhique du Gandhâra. 3 vols (Paris-Hanoi: E. Leroux, 1905–51), 130–142; A.
D.H. Bivar, ―Hāritī and the Chronology of the Kuṣāṇas,‖ Bulletin of the School of Orien-
tal and African Studies 33.1 (1970): 10–21; Zwalf, A Catalogue of the Gandhāra Sculp-
ture, I, 44, 116; II, pl. 92.
36
Mario Bussagli, Central Asian Painting (Geneva: Skira, 1979), 54, 58; Williams,
―The Iconography of Khotanese Painting,‖ 138–139.
37
On the iconography of Vārāhī in Indian art, see Thomas Eugene Donaldson, ―Oris-
san Images of Vārāhī, Oḍḍiyāna Mārīcī, and Related Sow-Faced Goddesses,‖ Artibus
Asiae 55.1–2 (1995): 155–182; on the Saptamātṛkās: Shivaji K. Panikkar, Saptamātṛkā
Worship and Sculptures: An Iconological Interpretation of Conflicts and Resolutions in
the Storied Brāhmanical Icons (New Delhi: D. K. Printworld, 2000).
38
Marc Aurel Stein, Ancient Khotan. Detailed Report of Archaeological Explorations
in Chinese Turkestan, 2 vols (Oxford: Clarendon Press, 1907), I, 278–279, II, pl. LX;
Williams, ―The Iconography of Khotanese Painting,‖ 142, fig. 51; Whitfield, Roderick,
The Art of Central Asia. The Stein Collection in the British Museum, 3. Textiles, Sculpture
and Other Arts (Tokyo: Kodansha, 1985), 316, pl. 70. On Maheśvara in the Khotan oasis,
see Williams, ―The Iconography of Khotanese Painting,‖ 142–145.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

27
Figure 7. Painting on the northern wall of temple D13. Dandan Oiliq. Author‘s draw-
ing after Baumer, ―Dandan Oilik Revisited,‖ fig. 24.

Figure 8. Vārāhī. Relief from Malagaon (Sirohi, Rajasthan), 8th–9th c. Ajmer, Raj-
putana Museum. Courtesy American Institute of Indian Studies.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

28
Maheśvara is flanked by a female figure with a swaddled child in her
lap, followed by a three-headed, four-armed ithyphallic god wearing a
dhotī, a necklace, two threads across his chest, with spiders clinging on
them, and snakes coiling around his arms. His central face is fearsome,
and the three circles on the nimbus probably indicate human skulls. His
attributes are sun and moon, and a trident held horizontally, with the
prongs towards the right. It seems that we deal with a local version of
Mahākāla, one of Maheśvara‘s emissaries, portrayed here more or less as
he is described in Buddhist sources. Our image can be compared with a
sculpture from Fattegarh (Kashmir), which—according to Phyllis Gran-
off—depicts a three-headed ‗Buddhist‘ Maheśvara and, on the reverse, a
gruesome Mahākāla, who has much in common with the Dandān-öiliq
deity: round open eyes, a human skull on the forehead, necklace and
yajñopavīta made of knotted snakes, a trident held horizontally; 39 his
proximity to Maheśvara, in our painting, is also remarkable.

____________
39
Phyllis Granoff, ―Maheśvara/Mahākāla: A Unique Buddhist Image from Kaśmīr,‖
Artibus Asiae 41.1 (1979): 64–82, fig. 1–3.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

29
Figure 9. Two deities, eastern corridor of temple CD4. Dandān-öiliq.
Author‘s drawing after Matsumoto, ed., Treasures of the Silk Road, fig. 32.

Part of a similar group of deities (including Skanda), arranged in an L-

shaped sequence framing larger images of seated and standing buddhas,
is preserved in the fragments of painting unearthed in 2002 by a Sino-
Japanese archaeological mission in the Dandān-öiliq temple named
CD4.40 On top of the vertical row (fig. 9), we see a female animal-headed
figure with an elongated attribute in her right hand. If the long, curved
muzzle is to be interpreted as an elephant trunk, as I think, we may iden-
tify this deity as a Central Asian Vināyakī, Gaṇeśa‘s female personifica-

____________
40
Zhang, Qu and Liu, ―A Newly Discovered Buddhist Temple,‖ 158, fig. 5, color plate
5.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

30
tion, also known as Gaṇeśvarī, Gajānanā, and other epithets.41 Vināyakī
is a late addition of the Saptamātṛkā sets (8th to 9th centuries), but she is
more at home among the 64 Yoginīs. The attribute she holds in her right
hand could therefore be a radish (Skt. mūlakakanda), which is held by
Gaṇeśa in his Khotanese representations42 as well as in the painting in
Dunhuang Mogao Cave 285 (fig. 4). The animal seated below her can
hardly be identified as a rat, Gaṇeśa‘s vāhana, but it curiously recalls the
vāhana on which the Vināyakī portrayed in the temple of the 64 yoginīs
at Hirapur (Orissa) stands, in which some scholars see a donkey43 (fig.
10). In the Dandān-öiliq mural, however, what seem to be the remains of
a small rat leaning to the figure‘s left knee are preserved.44
Below the elephant-headed goddess are a female figure with a swad-
dled child in her lap followed by a figure too poorly preserved to be
described.

____________
41
Prithvi Kumar Agrawala, Goddess Vināyakī. The Female Gaṇeśa (Varanasi: Prithivi
Prakashan, 1978).
42
Williams, ―The Iconography of Khotanese Painting,‖ 145–147, figs. 39, 54–56;
Whitfield, The Art of Central Asia, 309, 311, nos. 53.2, 57.
43
Cf., for instance, Thomas Eugene Donaldson, Tantric and Śākta Art in Orissa, II
(New Delhi: D.K. Printworld, 2002), 667. I am grateful to Chiara Policardi for pointing
out this evidence to me.
44
This detail escaped my attention in my former descriptions of this mural (Lo Muzio,
―Culti brahmanici a Khotan,‖ and Lo Muzio, ―Skanda and the Mothers‖).

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

31
Figure 10. Vināyakī. Relief in the Temple of the 64 yoginīs. Hirapur, Orissa. Courte-
sy Nilesh Korgaokar.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

32
In the horizontal row is a goat- or ram-headed four-armed female figure,
with sun and moon and an unclear object in her right hand (fig. 11), pos-
sibly a female counterpart of Naigameśa, a demon linked with children‘s
health; she is followed by a frowning Skanda, sitting on a peacock,
three-headed, four-armed, holding sun and moon, and a cockerel in his
right hand; black dashes are scattered on his skin. The last figure is a
frowning female deity with drooping breasts who lifts an emaciated na-
ked child, apparently ill or dead, holding him by the wrists and ankles.

Figure 11. Three deities, eastern corridor of temple CD4. Dandān-öiliq. Author‘s
drawing after Matsumoto, ed., Treasures of the Silk Road, fig. 32.

This overview ends with the mural painting found in the Dandān-öiliq
temple CD10, excavated in 2004 by the Sino-Japanese expedition, show-
ing a sequence of figures seated in three-quarter view to the left (fig.
12).45 A three-headed and four-armed Skanda, riding his peacock, hold-
ing the sun and moon emblems in his upper hands and a cock in the low-
er left, appears on the right of a row of seven female deities. The latter
are portrayed in the same attire, although in different colours, in
añjalimudrā (except the seventh from the right, with a swaddled child in
her lap); the lack of sharp iconographic differences makes these figures
almost interchangeable among themselves. On the left of the goddesses
____________
45
Dandan wulike yizhi, pl. 9. I thank Erika Forte for providing me with a picture of
this mural.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

33
is a cuirassed male holding a spear (Vaiśravaṇa?); further left, traces of a
female (?) figure are still visible, whereas the head of a female donor
holding a lotus flower is preserved in the register below (beneath the
cuirassed personage).

Figure 12. A row of female and male deities from temple CD10. Dandān-öiliq.
Author‘s drawing after Dandan wulike yizhi, pl. 9.

It seems clear that the recurrent representation of Skanda along with

female deities, either animal-headed or holding a child in their laps, links
all the Dandān-öiliq paintings I have been describing so far to the cult
addressed to ‗seizers, graspers‘ (Skt. grahas) or ‗children graspers‘ (Skt.
bālagrahas), documented in the epics (mainly in the Mahābhārata) and
in medical literature. Before being promoted to the rank of fierce warrior
and, then, of senāpati—a crucial redefinition which took place in the
Mahābhārata (Āraṇyakaparvan and Śalyaparvan sections), but not in
medical texts—Skanda was worshipped as a major spirit deity procuring
illness or death to children; he was believed to lead flocks of malign,
mostly female demoniac beings named grahāṇī, mātṛ or mātṛkā.46 Both
in the Mahābhārata and in Ayurvedic sources their description is usually
sketchy, and the many different lists of grahas recorded in the sources
____________
46
The topic is thoroughly treated by Richard D. Mann (Mann, The Rise of Mahāsena,
49–77), to whom I also refer the reader for a full account of written sources. See also
Dominik Wujastyk, ―Miscarriages of Justice: Demonic Vengeance in Classical Indian
medicine,‖ in Religion, Health and Suffering, ed. John R. Hinnells and Roy Porter (Lon-
don: Routledge, 2011), 256-275.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

34
are at variance among themselves as to the number of these spirits, their
names, and respective appearance (when noted). Yet it is not difficult to
see a reflection of the belief in such demons in a number of reliefs from
Kushan Mathurā (2nd to 3rd centuries), in which Skanda, in his martial
attire, appears along with ‗mothers‘, some of whom animal-headed (fig.
13). The architectural and cultic contexts which these reliefs belonged to
are unknown, but I would not rule out the possibility that they originate
from some Buddhist sacred area. Terracotta figurines depicting animal-
headed mothers with a child in their laps have been found in Kushan and
Gupta layers in several early Buddhist monasteries in Northern India
(Rājghāā, Kumrahār, Jetavana);47 to these, we can add the similar speci-
mens from Mathurā, although from unrecorded archaeological contexts.48

Figure 13. Relief showing the god Skanda beside a group of ‗mothers.‘ Mathurā.
Mathurā Government Museum. Courtesy American Institute of Indian Studies.

____________
47
Gethin Reesand Fumitaka Yoneda, ―Celibate Monks and Foetus-Stealing Gods:
Buddhism and Pregnancy at the Jetavana Monastery, Shravasti, India,‖ World Archaeolo-
gy 45.2 (2013): 255–265.
48
Panikkar, Saptamātṛkā orship and Sculptures, 31–53, pls. 9–13, 15.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

35
As I mentioned, Skanda and the grahas are very well-known in ancient
medical texts, such as the Suśrutasaṃhitā which covers a long time
span, from the Maurya to the Gupta epoch and the Kāśyapasaṃhitā (7th
century). 49 These sources seem to be unaware of or unaffected by the
reinterpretation of Skanda occurred in the Mahābhārata: in the Ayurve-
dic literature, Skanda remains a demon afflicting children.
That traditional medicine was part of the cultural heritage of Bud-
dhism since the earliest phase of its history is a well-established fact.50
Buddhist monasteries were important seats of medical education, and
Buddhist proselytism was a major factor in the spread of Indian medicine
towards Central Asia, the Far East, and Southeast Asia. Medical texts
from the Ayurvedic tradition were translated from Pāli and Sanskrit into
Chinese, Tibetan, as well as Khotanese; in this language, in particular,
two incomplete medical texts are preserved, the Siddhasāra by Ravigup-
ta (c. 650), translated from the Sanskrit and the Tibetan, and the so-
called Jīvakapustaka, otherwise undocumented. 51 Such texts, however,
need not be intended as a direct source for the Dandān-öiliq paintings. In
other words, it was not by the initiative of Central Asian monks or artists
that malevolent spirits drawn from Brahmanical medical sources were
‗pasted into‘ a Buddhist context, as we know that the grahas had already
been metabolized in the Buddhist ritual through the dhāraṇīs, or mystic
spells aimed at protecting from demons, diseases, beasts, hostile planets,
and other dangers. We have an outstanding piece of iconographic evi-
dence of this practice in the three extant folios of a
Mahāsāhasrapramardanī manuscript from Dunhuang, recovered by
Aurel Stein in Dunhuang, now kept in the British Museum.52 The original
manuscript (9th century) was composed of eight or nine folios. The
Mahāsāhasrapramardanī is one of the dhāraṇīs which were grouped in
____________
49
Mann, The Rise of Mahāsena, 31.
50
Kenneth G. Zysk, Asceticism and Healing in Ancient India. Medicine in the Bud-
dhist Monastery (Delhi: Motilal Banarsidass Publishers, 2000), 38–39.
51
Mauro Maggi, ―Khotanese Literature,‖ in The Literature of Pre-Islamic Iran (A His-
tory of Persian Literature, XVII), ed. Ronald E. Emmerick, and Maria Macuch (London:
I. B. Tauris & Co. Ltd., 2009), 413–415.
52
Whitfield, The Art of Central Asia, pl. 75. For a thorough philological analysis of
the Khotanese text, see Mauro Maggi, ―A Chinese-Khotanese Excerpt from the
Mahāsāhasrapramardanī,‖ in La Persia e l’Asia Centrale da Alessandro al X secolo (Ro-
me: Accademia Nazionale dei Lincei, 1996), 123–137.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

36
the so-called Pañcarakṣa, a pentad of spells, around the 12th century; the
single dhāraṇīs, however, are dated much earlier than that.53 Each leaf of
the Dunhuang manuscript presents the depiction of a female, animal-
headed demon on both sides, along with one or more children, accompa-
nied by a Chinese text with an abridged Khotanese translation, providing
the Khotanese and Sanskrit names of the figures and the respective dis-
ease they cause. One of the figures, an owl-headed demon here named
Mukhamaṇḍikā (Khot. Mukhamaṃdā), lifts a naked child, holding him
by the wrists and ankles, just like one of the ‗mothers‘ represented in
Dandān-öiliq temple CD4 (fig. 11).54

5. Conclusion

The significance of the Dandān-öiliq paintings for the issue I have been
dealing with does not need to be emphasized further. In the context of
the debate on Buddhism versus Brahmanism, summarised in my prelim-
inary remarks, the iconographic theme we have met in several Khotanese
paintings—Skanda as a leader of grahas or ‗mothers‘—is an archaic one
(unattested in Indian iconography after the Kushan epoch) and, as such,
is probably intended as part of a common cultural heritage shared by
Brahmanism and Buddhism (i.e. Seyfort Ruegg‘s ‗substratum‘). The way
in which these beliefs were embedded into Buddhist ritual practice was,
as we have seen, the dharaṇī. As to when this happened, the Chinese
translation of the Mahāmāyūrī, one of the dharaṇīs which were later
collected into the Pañcarakṣa, is dated to 317.55 I wonder whether this
may serve as a chronological indication for other dharaṇīs as well (like
the Mahāsāhasrapramardanī).

____________
53
Gerd J.R. Mevissen, ―Transmission of Iconographic Traditions: Paðcarakṣa Heading
North,‖ in South Asian Archaeology 1989: Papers from the Tenth International Confer-
ence of South Asian Archaeologists in Western Europe, Musée national des Arts asi-
atiques-Guimet, Paris, France, 3–7 July 1989, ed. Catherine Jarrige, with the assistance of
J. P. Gerry, and R. H. Meadow (Madison Wisconsin: Prehistory Press, 1992), 415.
54
For a female figure in a similar attitude in a painting from Khadalik (Khotan), see
Marc Aurel Stein, Serindia. Detailed Report of Explorations in Central Asia and West-
ernmost China, 5 vols (Oxford: Clarendon Press, 1921), pl. XI; Williams, ―The Iconogra-
phy of Khotanese Painting,‖ 139–140, fig. 45.
55
Mevissen, ―Transmission of Iconographic Traditions‖, 415.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

37
 Single elements of the Khotanese painted compositions, however, pa-
tently betray later borrowings (or updatings). Apart from the very image
of Skanda recorded in Dandān-öiliq as well as in all Central Asian and
Chinese depictions, drawn from the Brahmanical iconographic lexicon of
the Gupta period, some of the female characters of the Dandān-öiliq
paintings—in particular, Vārāhī and Vināyakī (if I have correctly identi-
fied them)—are more at home among the Saptamātṛkās (or even the
yoginīs) than in a row of grahas. For this very reason, they are to be
considered later additions, as the Saptamātṛkās are a conceptual and
iconographic Brahmanical ‗invention‘ which can be dated, broadly
speaking, to the early to mid-Gupta epoch, i.e. the 4th to 5th centuries
(the yoginīs are an even later phenomenon). It is particularly meaningful,
in this regard, that in the Mañjuśrīmūlakalpa (8th century?56) we find a
long and updated list of ‗mothers‘, including goddesses that we are ac-
customed to seeing in Saptamātṛkā sets, such as Brahmāṇī, Māheśvarī or
Vārāhī, as well as ancient grahāṇīs, such as Mukhamaṇḍitikā (wrongly
transcribed as Sukhamaṇḍitikā), Śakunī, or Pūtanā.57
My final remark on the intriguing iconographic and religious ‗stratig-
raphy‘ testified in the Khotanese murals I have been dealing with con-
cerns the inclusion, in Dandān-öiliq D13, of Maheśvara and (what I think
could be) Mahākāla (fig. 7). With the advent of the Saptamātṛkās, and
with the cleansing of his demoniac traits, Skanda gradually abdicates his
role of leader of the ‗mothers‘ to Śiva (Maheśvara); this transition took
place during the 6th century.58 As a leader of Saptamātṛkā groups, Śiva is
often accompanied by Mahākāla.59
Future findings in Central Asia, in Gandhāra as well as in the Indian
motherland, will surely improve our understanding of the dialogue be-
tween Buddhism and Brahmanism. For the moment being, it seems clear
that against the background of a common cultural heritage (whose actual
relevance should be cautiously evaluated case by case), Buddhists con-
stantly kept an eye on the Brahmanical religious system, literature, arts,
____________
56
Anthony K. Warder, Indian Buddhism (Delhi: Motilal Banarsidass Publishers,
1970), 525.
57
Āryamañjuśrīmūlakalpa I.21. I thank Dominic Goodall for providing me this refer-
ence.
58
Mann, The Rise of Mahāsena, 211.
59
Panikkar, Saptamātṛkā orship and Sculptures, 116, 120, 148, 175.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

38
and to the changes occurring within them through the centuries. There is
no other way, in my opinion, to explain the fascinating stratification of
borrowings we have discerned in the Central Asian iconographic reper-
toire of Skanda.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

39
 Abbreviations

Dandan wulike yizhi. Zhongguo Xinjiang wenwu kaogu yanjiusuo 中国新疆

文物考古研究所 and Riben Fojiao daxue Niya yizhi
xueshu yanjiu jigou 日本佛教大学尼雅遗址学术研究
机构, eds., Dandan wulike yizhi — Zhong-Ri gongtong
kaochayanjiu baogao 丹丹乌里克遗址. 中日共同考
察研究报告. Dandan Oilik Site. Report of the Sino-
Japanese Joint Expedition. Beijing: Wenwu chubanshe,
2009.

Bibliography

Primary Sources
Āryamañjuśrīmūlakalpa. In The Āryamañjuśrīmūlakalpa, edited by T. Gaṇapati
Śāstrī, Parts I-III. Trivandrum: Superintendent Government Press, 1920–1925.

Secondary Sources
Agrawala, Prithvi Kumar. Goddess Vināyakī. The Female Gaṇeśa. Varanasi: Prithivi
Prakashan, 1978.
Agrawala, R. C. ―Gandhāra Skanda with Flames.‖ East and West 18.1–2 (1968):
163–165.
Agrawala, R. C. ―Skanda with Cock and Peacock: A Rare Device in Gandharan Art.‖
In Eastern Approaches: Essays on Asian Art and Archaeology, edited by Thomas
S. Maxwell, 130–132. Delhi: Oxford University Press, 1992.
Agrawala, R. C. ―Skanda-Ṣaṣṭhī in Gandharan Reliefs—A Review.‖ East and West
45 (1995): 329–332.
Baumer, Christoph. ―Dandan Oilik Revisited: New Findings a Century Later.‖ Ori-
ental Art 49.2 (1999): 2–14.
Baumer, Christoph. Southern Silk Road: In the Footsteps of Sir Aurel Stein and Sven
Hedin. Bangkok: White Orchid Books, 2000.
Baumer, Christoph. ―Sogdian or Indian Iconography and Religious Influences in
Dandan-Uiliq: The Murals of Buddhist Temple D 13.‖ in The Art of Central Asia
and the Indian Subcontinent, edited by Anupa Pande, 170–184. New Delhi: Aryan
Book International, 2009.
Beyer, Stephan. The Cult of Tārā: Magic and Ritual in Tibet. Berkeley: University of
California Press, 1973.
Bhattacharya, Benoytosh. The Indian Buddhist Iconography, Mainly Based on the
Sādhanamālā and Cognate Tāntric Texts of Rituals. Calcutta: K. L. Mukhopadh-
yay, 1958.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

40
Bivar, A. D. H. ―Hāritī and the Chronology of the Kuṣāṇas.‖ Bulletin of the School of
Oriental and African Studies 33.1 (1970): 10–21.
Bronkhorst, Johannes. Buddhism in the Shadow of Brahmanism. Leiden: Brill, 2011.
Bussagli, Mario. Central Asian Painting. Geneva: Skira, 1979.
Davidson, Ronald. ―Review: The Symbiosis of Buddhism with Brahman-
ism/Hinduism in South Asia and of Buddhism with ‘Local Cults’ in Tibet and the
Himalayan Region by David Seyfort Ruegg.‖ Journal of the American Oriental
Society 129.1 (2009): 115–117.
DeCaroli, Robert. Haunting the Buddha. Indian Popular Religions and the For-
mation of Buddhism. Oxford: Oxford University Press, 2004.
Donaldson, Thomas Eugene. ―Orissan Images of Vārāhī, Oḍḍiyāna Mārīcī, and
Related Sow-Faced Goddesses.‖ Artibus Asiae 55.1–2 (1995): 155–182.
Donaldson, Thomas Eugene. Tantric and Śākta Art in Orissa, New Delhi: D.K.
Printworld, 2002.
Dunhuang wenwu yanjiusuo 敦煌文物研究所, ed. Dunhuang Mogao ku 敦煌莫高
窟 [Mogao Caves at Dunhuang], I. Beijing: Wenwu chubanshe, 1982.
Filigenzi, Anna. ―Maître du ciel, héros de la terre: la triade Bouddha–Vajrapāṇi–
Skanda dans l‘art du Gandhāra.‖ In Art et archéologie des monastères gréco-
bouddhiques du Nord-Ouest de l’Inde et de l’Asie centrale. Actes du colloque
international du CRPOGA, Strasbourg, 17–18 mars 2000, edited by Zemaryalai
Tarzi and Denyse Vaillancourt, 93–111. Paris: De Boccard, 2005.
Foucher, Alfred. L’art gr co-bouddhique du Gandhâra. 3 vols. Paris: E. Leroux,
1905–51.
Granoff, Phyllis. ―Maheśvara/Mahākāla: A Unique Buddhist Image from Kaśmīr.‖
Artibus Asiae 41.1 (1979): 64–82.
Harrison, Paul and Christian Luczanits. ―New Light on (and from) the Muhammad
Nari Stele.‖ In Special International Symposium on Pure Land Buddhism, 4th Au-
gust 2011, Otani University, 69–127. Kyoto: Ryukoku University Research Cen-
ter for Buddhist Cultures in Asia, 2012.
Joshi, Nilakanth Purushottam. Catalogue of the Brahmanical Sculptures in the State
Museum, Lucknow (Part I). Lucknow: The Archana Printing Press, 1972.
Lo Muzio, Ciro. ―Culti brahmanici a Khotan: note sulle pitture del tempio D13 a
Dandan Oiliq.‖ Rivista degli Studi Orientali (n.s.) 79.1–4 (2007): 185–201.
Lo Muzio, Ciro. ―Skanda and the Mothers in Khotanese Buddhist Painting.‖ In In-
teraction in the Himalayas and Central Asia: Processes of Transfer, Translation
and Transformation in Art, Archaeology, Religion and Polity, edited by Eva Al-
linger, Frantz Grenet, Christian Jahoda, Maria-Katharina Lang, and Anne Vergati,
71–89. Vienna: Österreichische Akademie der Wissenschaften, 2017.
Maggi, Mauro. ―A Chinese-Khotanese Excerpt from the Mahāsāhasrapramardanī.‖
In La Persia e l’Asia centrale da Alessandro al X secolo, 123–137. Rome: Acca-
demia Nazionale dei Lincei, 1996.
Maggi, Mauro. ―Khotanese Literature.‖ In The Literature of Pre-Islamic Iran (A
History of Persian Literature, XVII), edited by Ronald E. Emmerick and Maria
Macuch, 330–417. London: I. B. Tauris & Co. Ltd., 2009.
de Mallmann, Marie-Thérèse. Introduction à l’iconographie du Tântrisme
bouddhique. Paris: Librairie d‘Amérique et d‘Orient Adrien Maisonneuve, 1975.
Mann, Richard D. The Rise of Mahāsena. The Transformation of Skanda-Kārttikeya
in North India from the Kuṣāṇa to Gupta Empires. Leiden: Brill, 2012.
Mevissen, Gerd J. R. ―Transmission of Iconographic Traditions: Paðcarakṣa Heading
North.‖ In South Asian Archaeology 1989: Papers from the Tenth International

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

41
 Conference of South Asian Archaeologists in Western Europe, Musée national des
arts asiatiques-Guimet, Paris, France, 3–7 July 1989, edited by Catherine Jarrige,
with the assistance of J. P. Gerry and R. H. Meadow, 415–424. Madison Wiscon-
sin: Prehistory Press, 1992.
Mizuno, Seichi and Nagahiro, Toshio. Yün-kang. The Buddhist Cave-temples of the
Fifth Century A.D. in North China. Detailed Report of the Archaeological Survey
Carried Out by the Mission of the Tōhō Bunka Kenkyūsho 1938–45, 18 vols. Kyo-
to: Jimbunkagaku Kenkyūsho Kyoto University, 1951–55.
Mukherjee, Benoytosh. ―The Development of the Iconography of Kārttikeya
(Mahāsena) in Central Asia.‖ In Investigating Indian Art. Proceedings of a Sym-
posium on the Development of Early Buddhist and Hindu Iconography, held at the
Museum of Indian Art, Berlin, May 1986, edited by Marianne Yaldiz and Wibke
Lobo, 249–259. Berlin: Museum für Indische Kunst, 1987.
Pal, Pratapaditya. ―Sculptures from South India in The Art Institute of Chicago.‖ Art
Institute of Chicago Museum Studies 22.1 (1996): 20–35, 94.
Panikkar, Shivaji K. Saptamātṛkā orship and Sculptures: An Iconological Interpre-
tation of Conflicts and Resolutions in the Storied Brāhmanical Icons. New Delhi:
D. K. Printworld, 2000.
Pelliot, Paul. Les grottes de Touen-houang. Peintures et sculptures bouddhiques des
poques des ei des T’ang et des Song. Paris: Geuthner, 1920-1924.
Quagliotti, Anna Maria. ―Il Buddha che insegna la Legge: una stele raffigurante una
‗Terra Pura‘.‖ In Studi in onore di Umberto Scerrato per il suo settantacinquesi-
mo compleanno, edited by Maria Vittoria Fontana and Bruno Genito, 2 vols, II,
641–656. Naples: Istituto Italiano per l‘Africa e l‘Oriente, 2003.
Rees, Gethin and Fumitaka Yoneda. ―Celibate Monks and Foetus-Stealing Gods:
Buddhism and Pregnancy at the Jetavana Monastery, Shravasti, India.‖ World Ar-
chaeology 45.2 (2013): 252–271.
Regamey, Constantin. ―Motifs vichnouites et śivaïtes dans le Kāraṇḍavyūha.‖ In
Études tibétaines dédiées à la mémoire de Marcelle Lalou, 411–432. Paris:
Librairie d‘Amérique et d‘Orient Adrien Maisonneuve, 1971.
Sanderson, Alexis. ―Vajrayāna: Origin and Function.‖ In Buddhism into the Year
2000: International Conference Proceedings, 87–102. Bangkok: Dhammakāya
Foundation, 1994.
Seyfort Ruegg, David. ―Sur les rapports entre le Bouddhisme et le ‗substrat religieux‘
indien et tibétain.‖ Journal Asiatique 252 (1964): 77–95.
Seyfort Ruegg, David. ―Review: Introduction à l’iconographie du Tântrisme
bouddhique by Marie-Thérèse de Mallmann.‖ Journal of the American Oriental
Society 98.4 (1978): 543–545.
Seyfort Ruegg, David. ―A Note on the Relationship between Buddhist and ‗Hindu‘
Divinities in Buddhist literature and Iconology: the Laukika/Lokottara Contrast
and the Notion of an Indian ‗Religious Substratum‘.‖ In Le parole e i marmi. Stu-
di in onore di Raniero Gnoli nel 70° compleanno, edited by Raffaele Torella,
735–742. Rome: Istituto Italiano per l‘Africa e l‘Oriente, 2001.
Seyfort Ruegg, David. The Symbiosis of Buddhism with Brahmanism/Hinduism in
South Asia and of Buddhism with ‘Local Cults’ in Tibet and the Himalayan Re-
gion. Vienna: Österreichische Akademie der Wissenschaften, 2008.
Soper, Alexander Coburn. Literary Evidence for Early Buddhist Art in China (Arti-
bus Asiae, Supplementum, 19). Ascona: Artibus Asiae Publishers, 1959.
Southworth, Kirsten. ―Skanda in Gandhāra: A Hindu God in a Buddhist Environ-
ment.‖ In Changing Forms and Cultural Identity: Religious and Secular

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

42
 Iconographies, 1. South Asian Archaeology and Art. Papers from the 20th confer-
ence of the European Association for South Asian Archaeology and Art held in
Vienna from 4th to 9th of July 2010, edited by Deborah Klimburg-Salter and Lin-
da Lojda, 71–75. Turnhout: Brepols, 2014.
Srinivasan, Doris Meth. ―Skanda/Kārttikeya in the Early Art of the Northwest.‖ Silk
Road Art and Archaeology 5 (1997–98): 233–268.
Stein, Marc Aurel. Ancient Khotan. Detailed Report of Archaeological Explorations
in Chinese Turkestan, 2 vols. Oxford: Clarendon Press, 1907.
Stein, Marc Aurel. Serindia. Detailed Report of Explorations in Central Asia and
Westernmost China, 5 vols. Oxford: Clarendon Press, 1921.
Taddei, Maurizio. ―Non-Buddhist Deities in Gandharan Art.‖ In Investigating Indian
Art. Proceedings of a Symposium on the Development of Early Buddhist and Hin-
du Iconography, held at the Museum of Indian Art, Berlin, May 1986, edited by
Marianne Yaldiz and Wibke Lobo, 349–362. Berlin: Museum für Indische Kunst,
1987.
Matsumoto, Nobuyuki, ed. Treasures of the Silk Road. Recent Discovery from Xin-
jiang and Shaanxi. Tokyo: Nippon Hōsō Kyōkai, 2005.
Warder, Anthony K. Indian Buddhism. Delhi: Motilal Banarsidass Publishers, 1970.
Whitfield, Roderick. The Art of Central Asia. The Stein Collection in the British
Museum, 3. Textiles, Sculpture and Other Arts. Tokyo: Kodansha, 1985.
Williams, Joanna. ―The Iconography of Khotanese Painting.‖ East and West 23.1–2
(1973): 109–154.
Wujastyk, Dominik. ―Miscarriages of Justice: Demonic Vengeance in Classical
Indian medicine.‖ In Religion, Health and Suffering, edited by John R. Hinnells
and Roy Porter, 256–275. London: Routledge, 2011.
Zhang Yuzhong, Qu Tao, and Liu Guorui. ―A Newly Discovered Buddhist Temple
and Wall Paintings at Dandan-Uiliq in Xinjiang.‖ Journal of Inner Asian Art and
Archaeology 3 (2008): 157–170.
Zhongguo Xinjiang wenwu kaogu yanjiusuo 中国新疆文物考古研究所 and Riben
Fojiao daxue Niya yizhi xueshu yanjiu jigou 日本佛教大学尼雅遗址学术研究机
构, eds., Dandan wulike yizhi — Zhong-Ri gongtong kaochayanjiu baogao 丹丹
乌里克遗址. 中日共同考察研究报告. Dandan Oilik Site. Report of the Sino-
Japanese Joint Expedition. Beijing: Wenwu chubanshe, 2009.
Zysk, Kenneth G., Asceticism and Healing in Ancient India. Medicine in the Bud-
dhist Monastery. Delhi: Motilal Banarsidass Publishers, 2000.
Zwalf, Wladimir. A Catalogue of the Gandhāra Sculpture in the British Museum, 2
vols. London: The Trustees of the British Museum, 1996.

BuddhistRoad Paper 6.1 Special Issue. Lo Muzio, ―Brahmanical Deities in Foreign

Land‖

43